SCANNING ELECTRON MICROSCOPE STUDY OF
CRACKS AND PORES IN CRYSTALLINE ROCKS
by
EVE SILVER SPRUNT
S.B.,
Massachusetts Institute of Technology
1972
Submitted in Partial Fulfillment
of the Requirements for the
Degree of Master of Science
at the
MASSACHUSETTS INSTITUTE OF TECHNOLOGY
.June,
Signature
of
1973
Author
............
Department of Earth and
Planetary Sciences
May 11, 1973
Certified by
Thesis Supervisor
Accepted by
Chairman, Departmental Committee
on Graduate Students
W
OWN
I,03
RAWN
IES
ABSTRACT
Scanning Electron Microscope Study of
Cracks and Pores in Crystalline Rocks
by
Eve Silver Sprunt
Submitted to the
Department of Earth and Planetary Sciences
on May 11, 1973
in partial fulfillment of the requirements
for the degree of Master of Science
Direct observation of the microstructures in rocks was
made possible through combined use of ion bombardment in sample
preparation and the scanning electron microscope for viewing.
Samples from six granites, a diabase, and a aabbro were
observed. The porosity can be divided into pore-like and
crack-like features. Pore concentrations are characteristic
of plagioclase in granite. Most of the crack-like features
are composed of rows of slots. For granites the aspect ratios
are on the order of 103 or.less; most cracks were 1 vim or less
wide; and 2 to 4 cracks frequently met at a common point.
A preliminary study of the effects of mechanical and
thermal stress was done, using two granites and a diabase that
had been stressed in triaxial compression to over 95% of their
fracture strength and a granite that had been heated to 400*F.
In the mechanically stressed granites the rows of slots
observed in the unstressed material were replaced by single
high aspect ratio cracks. The cracks tended to be oriented
parallel to the direction of maximum compression. In the
thermally stressed granite, existing cracks extended and more
grain boundaries were cracked. Quartz grains showed more new
cracks than the other materials.
Thesis Supervisor:
Title:
William F. Brace
Professor of Geology
TABLE OF CONTENTS
Page
Abstract ...............................................
i
Introduction ...........................................
1
History of the Problem ............................
Present Investigation .............................
2
Previous Attempts to Map Cracks in Rocks ..........
4
Method .................................................
7
Rocks Studied ..........................................
9
Observations - Unstressed Rocks ........................
10
General ...........................................
10
HARCs .............................................
Pores .............................................
12
13
Crack Intersections ...............................
14
Grain Boundaries ..................................
15
Observations - Mechanically Stressed Rocks .............
16
Specimens .........................................
Comparison with Unstressed Rocks ..................
16
Observations - Thermally Stressed Rocks ................
18
Discussion .............................................
Is the Section Really Undisturbed? ................
Comparison of Observation and Predictions .........
19
19
19
Scale of Features .................................
The Concept of Cracks and Pores ...................
24
Suggestions for Future Research ........................
27
Acknowledgements .......................................
30
I. Polishing Holder ......................
31
II. Choice of Coating .....................
III. Holder for the Rock ...................
32
Appendices:
IV. Precise Relocation of a Particular
SEM Field .............................
V. Microprobe Analysis ...................
16
25
33
34
36
TABLE.OF CONTENTS
Page
Bibliography .........
38
...........
...........
..........
List of Illustrations
List of Tables .......
Figures
43
.................
................
44
..............
45
Tables ...............
57
iii
INTRODUCTION
History of the Problem
Cracks and pores are thought to have important effects
on the physical properties of crystalline rocks [summarized,
for example, in Brace, 1972].
Compressibility, electrical
resistivity, thermal conductivity, and the velocity of
compression and shear waves exhibit large changes, which
have been attributed to the closure of cracks, over the
first one to two kilobars of effective pressure.
Adams and Williamson
[1923] were the first to suggest
that cracks and pores were responsible for the non-linearity
of elastic properties of rocks.
As cracks close in a
compressibility test, the bulk modulus increases and the
pressure-strain curve steepens.
After the first kilobar or
two of pressure the cracks are closed and the curve is linear
because the behavior of the rock is determined by constituent
minerals which are nearly perfectly elastic [Walsh and Brace,
1966; Walsh, 1965a; Brace, 1965].
Walsh derived expressions
for elastic constants by considering the effects of cracks
and pores [Walsh, 1965a, 1965b, 1965c; Walsh and Brace, 1966].
The cavities were modeled as isolated simple geometric shapes
in an infinite linearly elastic material.
In water-saturated crystalline rocks electrical
conduction is almost totally through pore fluids.
At low
pressure a drastic increase in electrical resistivity is
observed due to the closing of cracks.
Porosity is the sole
property which determines the high pressure resistivity of
water-saturated rocks composed of non-conducting minerals
[Brace and Orange, 1968; Orange, 1965].
The correlation
between electrical resistivity and porosity requires that
most of the cracks in rocks be interconnected.
Cracks form effective barriers to the flow of heat
when rocks are dry.
When the rock is saturated or when
cracks are closed by stress, the thermal conductivity
increases [Walsh and Decker, 1966].
Since saturation
increases the thermal conductivity, most of the cracks are
probably interconnected.
In dry rocks with porosity in the form of cracks, the
velocities of both compressional and shear waves at atmospheric
pressure are often significantly lower than at a few kilobars
of pressure.
When such rocks are saturated with water the
difference between compressional velocity at low pressure
and at high pressure is reduced.
The difference between dry
and saturated compressional velocities increases with
increasing crack porosity.
volumetric porosity.
There is no
correlation with
The shape of the cavities is important
because the compressional velocity is only increased when
the saturated cavities are in the form of cracks [Nur and
Simmons, 1969].
Porosity of crystalline rocks, including those with
low porosity can be measured with an accuracy of .001.
This
suggests that most cavities are interconnected [Brace, 1971b].
In a study of permeability of granite, Brace et al [1968] used
a model consisting of a single channel of constant crosssectional area.
Measured permeability and porosity determined
the hydraulic radius and allowed the shape of cavities in
granite to be estimated.
At low pressure, the cavities would
be flat slits, while at higher pressure they would be equidimensional pores.
What is the real form of the cavities?
interconnected or isolated?
and pores?
boundaries?
Are the cracks
How concentrated are the cracks
Where are they located relative to the grain
What kinds of aspect ratios do the cracks have?
These are just a few of the questions that studies of the
physical properties of rocks have raised.
Present Investigation
In this paper we show as reported earlier [Brace et al.,
1972] that direct observation of details of porosity and
grain structure in crystalline rocks is possible through
combined use of ion bombardment for sample preparation and
the scanning electron microscope for viewing.
Details of the
preparation of six granites, a diabase, and a gabbro are
given.
Cavity shapes, aspect ratios, connectivity and
location are reported and compared with predictions from
indirect measurements by other investigators.
A preliminary
study is made of mechanically and thermally stressed rocks
and the observed characteristics are compared with those of
the unstressed rocks.
Previous Attempts to Map Cracks in Rocks
Many attempts have been made to map cracks in rocks
using thin or polished sections.
In some instances an
effort was made to decorate the cracks before sectioning.
Koide and Hoshino [1967] investigated microfractures
in experimentally deformed rocks using thin sections.
Their
criteria for an open crack was one that appeared dark under
crossed nicols and clear under uncrossed nicols.
From their
study they obtained the preferred orientation of microcracks.
Willard and McWilliams [1969] determined the orientation of
defects in thin sections.
By their definition defects
include grain boundaries, cleavage planes, twin planes, slip
planes, inclusion planes, deformation bands, and microfractures.
Defect orientation analysis records all cracks irrespective
of their origin in time.
Paulding [1965] observed many
transgranular cracks in thin sections of undeformed Westerly
granite.
From thin and polished sections he reported that at
stress differences of about half the fracture strength cracks
began to grow from the pre-existing transgranular cracks and
grains and were inclined to the direction of maximum compression
and generally en echelon.
Crack growth continued as stress
was raised and was predominantly parallel to the direction of
5
maximum compression.
Wawersik and Brace [1971] used
polished sections to study crack densities, crack orientations,
crack location relative to grain boundaries and orientation
for the longer partial fractures in stressed rocks. One
difficulty they noted was in distinguishing cracks from grain
boundaries.
Baldridge and Simmons [1971] impregnated rocks with two
types of water-thin compounds.
One of the compounds used
Canada Balsam to mark the cracks with its yellow color.
The
other compound was a mixture of furfuryl alcohol, which easily
penetrated the pores, and hydrochloric acid, which polymerized
the alcohol.
The impregnated sample was heated to 750*C to
complete the polymerization of the alcohol and to convert
the polymer to carbon.
colored the cracks.
Sub-micron sized particles of carbon
After the impregnation procedures the
cracks were sectioned and polished.
Both compounds enabled
them to differentiate between original cracks, which were
decorated, and cracks introduced in the thin sectioning
process.
From their studies they were able to get crack
densities, crack orientations, average lengths, and crack
intersections.
Hadley [unpublished] attempted crack mapping
using radioactive tritiated furfuryl alcohol.
It was hoped
that the introduction of a radioactive polymer into the cracks
would permit a three-dimensional picture of the fissures to be
built up by progressively sectioning the rock and placing it
in contact with photographic film.
Unfortunately, film with
a fine enough grain size was not available.
Timur et al
[1971] used the scanning electron microscope
to observe samples that had been polished by a technique
similar to that used in making thin sections.
They were
primarily concerned with the structure of pores.
All of the above studies produced unreliable results
for one or more of the following reasons.
In the projects
based upon the observation of ordinary thin or polished
sections, it is difficult, if not impossible, to differentiate
between the original cracks and those created in the
sectioning process.
Often in polished sections it is a
matter of personal bias whether a feature is a crack or a
grain boundary.
Crack decoration schemes enabled positive
identification of some of the pre-sectioning cracks, but
failed to show cavities that had received none of the marking
fluid because they were disconnected from the outside.
No attempts were made to examine the fine details of
cavity structure.
Aspect ratios which depend on measurement
of both length and width were not given.
Of all the results reported from these studies the
crack orientations are probably the most accurate.
If the
cracks produced in the sectioning process were randomly
oriented, they would appear as background in the measurements
of the preferred orientation of the total crack population.
The total number of cracks would be increased, but the
preferred orientation would remain the same.
METHOD
Cores 1.04 cm in diameter were drilled from large
oriented blocks.
After the ends were ground perpendicular
to their axes, the cores varied from 1.14 to 1.27 cm in
length.
Along the axis of each core was cut a groove which
when viewed from one end served as an origin of coordinates.
The following procedures gradually eliminated the
damaged layers from the surface of the sample.
About 300
to 400 pm of rock was removed by hand grinding with
increasingly fine silicon carbide abrasives (240, 400, and
600 grit) on a metal wheel, followed by polishing with
aluminum oxide abrasives (5, 1, and .3 pm) on cloth laps.
The next step was the crucial one.
A layer about 30 to
40 pm thick was removed by ion bombardment (which is also
known as ion thinning) in a Commonwealth Scientific ion
milling instrument.
An ionized argon beam of about 6 kv
bombarded the sample at an angle of 170.
This ion thinning
was the vital difference betweenour specimens and the
ordinary thin or polished section.
Ion thinning does not produce a perfectly smooth
surface.
Olivine and orthopyroxene thin at about 1 pm/hr.
Plagioclase thins at 2 pm/hr.
Quartz thins at 3 to 8 pm/hr.
Mica thins even faster than quartz.
Thinning time was kept
close to the minimum period necessary to remove the damaged
layer (six hours for most rocks) so as to avoid very large
topographic steps between different minerals.
Bombardment
also tends to produce gentle mounds much smaller than the
grain size.
These mounds appear to be a function of the
beam rather than the mineralogy.
Areas of weakness such as
the regions surrounding cracks and pores do not thin faster
than the rest of the mineral.
Grain boundaries are not etched
out.
Sinde specimens viewed with a scanning electron
microscope (SEM) must have conducting surfaces, approximately
400 A of a 40% gold 60% palladium alloy was vacuum evaporated
onto the surface.
There are many good references on the use
of the SEM [Kimoto and Russ, 1969; Majumdar, 1970; Oatley et
al., 1965].
To get a three-dimensional picture of the porosity,
some samples were serially sectioned after photographs had
been taken of selected fields under the SEM.
First the metal
coating was removed with 1 and 0.3 ym aluminum oxide abrasives.
Then the surface was ion bombarded for six hours.
If polishing
did not precede thinning, any small irregularities in the
coating seemed to cause very uneven thinning of the rock.
The effects of thermal stress were studied in a sample
which, after being prepared as described above and photographed,
was placed for 31 hours in a furnace at 4004F, which was
calculated to be more than sufficient to induce many cracks
in the rock.
The sample was allowed to cool inside the
furnace overnight.
Recoating was unnecessary and the sample
could be viewed immediately after removal from the furnace.
The mechanically stressed samples came from triaxial
compression tests of dogbone-shaped specimens.
The rocks
were recovered intact after being loaded to over 95% of
their fracture strength [Brace, Paulding
Paulding, 1965].
and Scholz, 1966;
The dogbones were cast in a mold, sawed in
half lengthwise, and cores were drilled from their centers.
The cores surfaces were polished, thinned, and coated, as
described above.
ROCKS STUDIED
The rocks arranged in order of increasing total
porosity are listed in Table 1, along with their crack and
pore porosities, grain size, and modal analysis.
The
Westerly granite is from a block of so-called Westerly granite
referred to in Birch [1960].
Many experiments have been
conducted on other blocks of similar granite from Westerly,
including permeability [Brace, Walsh and Frangos, 1968],
linear compressibility [Brace, 1965], resistivity [Brace and
Orange, 1968], shear wave velocity [Simmons, 1964], and
compressional wave velocity [Birch, 1960].
Maryland diabase
is from Block I [Birch, 1960] and was used in many of the
same studies as Westerly granite [Brace, 1965; Brace and
Orange, 1968; Simmons, 1964; Birch, 1960].
The Chelmsford
granite is similar to that studied by Todd and Simmons [1972]
and Birch [1960].
San Marcos gabbro was used by Brace and
Orange [1968] and Birch [1960].
Raymond granite from the
Raymond Quarry in Madera, California, granodiorite of Red
Lake (referred to as Red Lake granite) near Shaver Lake,
California, and two
samples of Katahdin granite from Baxter
State Park, Maine, were also studied.
OBSERVATIONS - UNSTRESSED ROCKS
General
Our samples when viewed with an SEM look radically
different from thin sections of the same rocks viewed with
an optical transmission microscope.
SEM fields resemble
fields seen with a reflecting optical microscope.
With both
the SEM and the reflecting microscope, only surface structures
can be seen.
An important difference in the capabilities of
the two microscopes is the depth of focus; for the SEM it is
30 to 40 times greater than for the light microscope at a
given magnification [White and Roy, 1968].
The other difference
is in the range of magnifications available; an SEM is capable
of magnifying an object from 100 to 30,000 times.
In SEM micrographs, cavity rims and sharp topographic
steps frequently appear bright, because in these areas the
11
secondary electrons have an unusually large surface through
Dark regions such as the centers of
which they can escape.
cracks may result from the shadowing of one region by another
from the incident beam, or from the failure of electrons from
a depressed region to reach the collector [Majumdar,
1970;
Oatley et al., 1965; Kimoto and Russ, 1969].
As mentioned earlier, ion bombardment tends to produce
gentle mounds, 40 ym or less in diameter, which are a function
of the beam.
In Figure 6 the cracks near "c" are seen to have
no effect on the mounds.
Mineralogy was determined from structural details such
as cleavage or characteristic crystal form, or from composition
determined with the electron microprobe.
When a rock is viewed at low magnification, the cavities
seem to fall into two classes, pore-like and crack-like.
The
pore-like cavities are usually equant, but may have been aspect
ratios as high as 10.
Frequently they are randomly oriented
as in plagioclase (Figure 2),
(Figure 3).
but they also appear in lines
At high magnifications the crack-like features
often resolve into rows of slots.
A slot is a blunt-ended
cavity of nearly rectangular cross-section.
Many slots have
aspect ratios greater than 10 and may be referred to by the
general term, high aspect ratio cavity (HARC).
Other crack-
like features resolve into true cracks, which are HARCs with
sharp ends.
HARCs
Long, narrow, sharp-ended cracks as imagined in t
are relatively rare in many rocks.
The term HARC is u
to avoid some of the ambiguity and implicit shape invo.
the use of the term crack to refer to individual caviti
The use of the word crack is reserved for sharp-ended
cavities and high aspect ratio features that appear cracklike at low magnification.
Figure 1 shows a typical crack intersection in Westerly
granite.
The slots have black interiors and white edges.
At
"a" and "b" are two of the many delicate bridges dividing the
cracks into slots.
At lower magnifications these bridges are
invisible, so these cracks seem to be uninterrupted HARCs.
Individual slots often have blunt ends as seen here.
In most granites cracks have widths of 1 im or less.
Along their length cavities vary in width, making it difficult
to determine their aspect ratio.
Measuring HARC length was
difficult because the many tiny bridges made it hard to
determine if a HARC were continuous and because the curves
made it difficult to measure total length.
The indicated
problem is treated in more detail when the comparison between
observed and predicted aspect ratios is discussed.
Most HARCs
are continuous for only 10 to 100 ym, which is much less than
a grain diameter.
However, cracks composed of many slots had
lengths on the order of a grain boundary in all of the granites
except Westerly.
Occasionally a mica grain was cracked for its
Extremely narrow HARCs, such as
entire length by a HARC.
those 0.1 pm wide, accounted for the highest aspect ratios,
The aspect ratios for rocks
which were on the order of 103.
with very low crack porosities such as Maryland diabase and
San Marcos gabbro were about the same, but the widest cracks
were about 0.33 pm wide.
Pores
Pore-like features appear in several configurations,
one of which is as scattered cavities.
Pores of this type
can be seen below point "b" in the quartz grain in Figure 2.
Some of these pores are connected by fine cracks and others
appear isolated in this section.
Dense concentrations of pores in feldspar, especially
plagioclase, are common in granite.
(an1 7 ), Chelmsford (an2 o- 2 5),
Three granites, Westerly
and Red Lake (an3 o) have
concentrations of pores 7 im or less in diameter, with many
pores about .4 ym in size.
In Katahdin granite I almost all
of the plagioclase pores are less than .7 ym in diameter,
while in Raymond granite many are about 10 pm in diameter.
In Chelmsford granite (Figure 2) plagioclase grains have pore
porosities of 2 to 3%.
The microcline in Chelmsford granite
has a high pore porosity, but not as high as the plagioclase.
In Westerly granite the plagioclase, as mentioned, contains
pore concentrations, but the microcline is relatively porefree.
The plagioclases in San Marcos gabbro (an4 2 ) and
Maryland diabase (an,7 ) are not sites of pore concentrations.
Pores also appear in linear arrays.
cracks connect these pores.
In some instances
Other times the pores are isolated
inclusions enclosing daughter crystals (Figure 3).
These
trains of pores may mark the site of a healed crack.
Crack Intersections
Isolated HARCs, often with blunt ends, are not uncommon,
but such HARCs may be connected with other cavities beneath the
surface.
Sometimes a thin band of uncracked rock lies between
two HARCs which at low magnification seem to connect as at
point "c" in Figure 1.
The number of cracks observed sharing a common point
of intersection is generally 2 to 4 in granite.
In Red Lake
granite, 6 cracks meeting seems to be an upper bound, while
in Westerly, 3 cracks meeting is maximum.
Sometimes it is difficult to determine whether a feature
consists of one or more cracks, because the feature is very
sinuous and/or is composed of slots.
Some of the problem can
be explained with reference to Figure 2.
At "a" the crack
strikes to the northeast, but at the plagioclase-quartz
boundary, the crack curves until it strikes almost due east.
Then it makes several sharp zig-zags until it curves gently
near "b" so that it again strikes to the northeast.
Since
most of the major direction changes occur at places with
relatively large radii of curvature, the entire feature was
considered a single crack, although aspect ratios were only
15
measured between bridges.
When a large change in direction
occurs at a sharp corner, the feature is generally considered
to be two cracks.
Crack counting tends to be subjective
because of the continuous spectrum of radii of curvature
possible at corners.
Grain Boundaries
The percentage of the grain boundary that is cracked
varies significantly from rock to rock.
In San Marcos gabbro
(Figure 3) and Maryland diabase, the grain boundaries seem to
be crack-free.
From 20 to 85% of each grain boundary is
cracked in Westerly granite.
Examination at high magnification
of the quartz grain in Westerly granite seen in Figure 4 shows
that about 20% of its grain boundary is cracked.
The HARCs
all have aspect ratios less than 100 and are located in the
bright regions designated "a", "b", and "c".
In Chelmsford,
Raymond and Katahdin I granites about 20 to 90% of each grain
boundary is cracked.
Almost all of the grain boundaries in
Red Lake and Katahdin II granites are cracked.
Even when
most of a grain boundary is cracked, the cracking usually
consists of many HARCs whose lengths are all much shorter than
the apparent grain diameter.
OBSERVATIONS - MECHANICALLY STRESSED ROCKS
Specimens
Two samples of Westerly granite and one of Maryland
diabase were made from cores that had been stressed in triaxial
compression to within a few percent of their fracture strength.
The strength of the rocks and the confining pressure are
listed in Table 2.
Comparison with Unstressed Rocks
Westerly granite:
The sharp contrast between stressed
and unstressed Westerly granite can easily be seen by comparing
Figure 4 and Figure 5.
In the stressed rock the grain boundary
is occupied by HARCs of lengths equivalent to a grain diameter.
At point "a" in Figure 5, a HARC with an aspect ratio of a
little more than 200 is located in the grain boundary.
The
HARCs are much broader and longer than in unstressed rock.
At point "c" in Figure 5 there are cracks running from the
grain boundary into the interiors of the grains on both sides.
This type of cracking, beginning at the grain boundary, is
typical of stressed Westerly granite but is never seen in the
unstressed rock.
Multiple crack intersections are common as
at points "b" and "c" in Figure 5.
In some places fine cracks
crisscross the rock.
Figure 6 shows some of the details of cracks in stressed
Westerly granite to be compared with Figure 1 for unstressed
material,
The delicate bridges are gone and the cracks are
composed of single HARCs and exhibit a high degree of interconnectivity.
Many of the cracks tend to be subparallel to
the direction of maximum compression, but in feldspar the cracks
are- often. at a large angle to the principle stress direction
because they follow cleavage planes.
The orientations of the cracks in Figure 7a were
measured and plotted in the histogram in Figure 7c.
Although
only 217 non-grain boundary cracks were plotted, a definite
peak is evident near zero degrees to the direction of maximum
compression.
In a similar sized field in unstressed Westerly
granite, there were only 83 cracks, and their scattered
angular distribution is indicated on the histogram for
comparison.
Maryland diabase:
Although this rock was stressed
nearly to fracture, most of the grain boundaries are uncracked.
Steep topographic steps exist at the grain boundaries due to
differences in thinning rates.
In Figure 8 a crack can be seen
crossing a steep plagioclase-pyroxene boundary which is intact.
Many intra-granular cracks are also visible.
Other fields show
that the mineral cleavage planes exert a strong influence on
intra-granular cracking.
18
OBSERVATIONS -
THERMALLY STRESSED ROCK
A sample of Chelmsford granite was heated to 400*F in a
furnace at atmospheric pressure.
The sample was heated for
31 hours and allowed to cool inside the furnace to room
temperature.
Before the thermal stressing, selected fields
in the surface had been thoroughly examined and photographed
with the SEM.
These same fields were re-examined and re-
photographed after the heat treatment.
The bridges across the cracks were broken.
cracks extended and intersected.
in width.
Existing
Some cracks more than doubled
Sharp ends were observed on many HARCs.
At some
sharp corners of cracks which were present before heating, new
cracks formed.
were opened.
Many previously uncracked grain boundaries
Within some of the large quartz grains there
were many new intersecting cracks.
Although the feldspars
are extremely porous (2.5%) the pores remained disconnected
in the plane of observation.
The new cracks were observed at places which would be
expected to be weak-grain boundaries, crack tips and corners.
More new cracking marred the quartz than the other minerals.
However, it is difficult to determine whether a crack is
really new or just an extension of a previously existing
crack, because we are limited to two dimensions.
The "new"
cracks could be extensions of cracks, which did not intersect
the plane of observation before thermal stressing.
DISCUSSION
Is the Section Really Undisturbed?
There are several arguments in favor of the hypothesis
that we are seeing cracks and poresas they exist in the
interior of the rock.
One argument is based on the
observation that only a few of the cracks appear to be of
purely brittle origin.
Cracks of brittle origin, such as
those seen in stressed rock would have sharp ends and would
usually intersect each other.
Instead, blunt-ended cracks
which stop short of intersection are frequently seen.
Delicate bridges, which could be due to healing since the
rock was cracked early in its history, span the cracks.
Paulding [1965] reported many transgranular cracks in thin
sections of unstressed Westerly granite, which we have never
seen.
More support for the success of our damage removal
technique comes from the fact that some rocks such as Maryland
diabase are virtually devoid of cracks.
Furthermore, the
crack porosity of Chelmsford granite determined in our study
is in close agreement with that measured by linear
compressibility experiments.
Comparison of Observations and Predictions
Brace [1964, 1961] and Brace and Walsh [1966] have
predicted that microcracks have lengths on the order of the
grain size.
This length estimate was based on the Griffith
theory for tensile strength.
20
T
T
0
=
[2Eyj 2
1C
= tensile strength
E = Young's modulus
y = specific surface energy
c = half the length of the Griffith crack
2c = the longest straight section of grain boundary
(maximum grain diameter)
HARCs with lengths equivalent to the grain diameter of
one of the moderate to large grains in the granites have not
been observed.
There are features of grain boundary size -in
some granites (Red Lake, Katahdin II and Chelmsford), but they
are either long series of en echelon slots or cracks which are
bridged every 20 to 30 pm.
These features could be cracks
which formed early in the rock's history and have since undergone varying amounts of recrystallization.
In some places the
bridges seem to be from recrystallization.
With time, such
bridged cracks might become a trail of pores as in Figure 3.
Another possibility is that the slots were formed by stresses
which were not large enough to create continuous cracks, such
as those seen in the laboratory stressed samples.
It is
difficult to determine how extensive the bridging must be
before the crack ceases to behave elastically like a single
crack.
A three-dimensional picture of the bridges is not
available, so we do not know if they are pillars or more
substantial walls.
Do they crumble at relatively low pressures,
or do they survive so that the aspect ratios of the slots
between them can be used in predicting elastic properties?
The longest continuous HARCs in granites are about 200 to
400 ym long.
Long, straight HARCs are often found in mica.
In most granites almost all cracks are less than 1 ym
wide.
The exception is Katahdin TI, which contains large
cracks about 100 pm wide.
Cracks in San Marcos gabbro and
Maryland diabase are much less common, and those observed
are less than .4 ym wide.
Brace [1971b], because of the
failure of attempts to obtain a detailed picture of crack
structure, predicted that cracks would be less than 10 ym
wide.
The aspect ratio (or ratio of crack length to crack width)
is frequently referred to in crack theory.
Walsh [1965b] and
Brace, Walsh and Frangos [1968] have predicted from the pressure
to close elliptical cracks that aspect r.atios should be on the
order of 103.
Problems arise in getting both the correct
length and the correct width of a crack for determining its
aspect ratio.
The cross-section of a HARC is frequently very
irregular, so that, for example, in some places it might be
1 ym wide and in others only .5 ym wide.
was used for our ratios.
The average width
For bridged cracks the lengths of
the individual HARCs were used, but conceivably if the bridging
were not too substantial the length of the entire row of slots
should have been used.
Can the aspect ratios measured be used in the formulas
derived in crack theory?
Elastic analyses invariably treat
the case of an isolated cavity in an infinite medium with
separation of one crack length being a good approximation to
infinity [Walsh and Brace, 1966].
Cracks are frequently much
closer together than one crack length.
Our aspect ratios for
granites have about the order of magnitude as those predicted,
103,
but a great many cracks have ratios less than 500.
Crack porosity can be determined to ±.05% from the
intercept on the strain axis of the linear portion of the
pressure-strain curve extrapolated to zero pressure [Walsh,
1965b].
Using this method, Todd [personal communication] found
the crack porosity of Chelmsford granite's rift face to be .29%.
By mapping cracks in a field of Chelmsford in the same direction,
we found the porosity to be .26%.
Lack of a pronounced low pressure toe in the resistivity
pressure curve indicates zero crack porosity [Brace and Orange,
1968].
The resistivity curve for Maryland diabase does not
have a low pressure toe.
In agreement, we found that the
crack porosity of Maryland diabase is negligible.
Pore porosity can be determined to ±.1% by immersion.
Total porosity is given by
s
b
S
Zs = mineral specific gravity
Zb
bulk specific gravity
The mineral specific gravity is found by immersion in carbon
tatrachloride.
Bulk specific gravity was measured in dried,
precisely-ground cylindrical samples.
Pore porosity is
found by subtracting the crack porosity from the total
porosity [Brace, 1965].
For Chelmsford granite the total
porosity was 1.0%, so the pore porosity is 0.7%
[Todd, personal
communication].
By determining the typical porosity in photomicrographs
of the major mineral phases in Chelmsford granite and
multiplying each porosity by the percentage of the rock it
comprises, a porosity of 1.2% was determined for that granite.
The high estimate could be due to the pores in the plagioclase.
Many of these pores are probably isolated fluid inclusions
that do not contribute to porosity as measured by immersion.
The range in size of the pores in plagioclase from
10 ym in diameter in Raymond to less than .7 ypm in diameter
in Katahdin I may be related to pressure and hence to the
depth at which the granite solidified.
On this basis, Katahdin
granite would be a shallow granite, whereas Raymond granite
would be a deep one.
Chelmsford, Westerly, and Red Lake would
lie somewhere between those extremes.
Another possibility is
that the pores are a function of whether or not the rock is
water-saturated.
A porosity-resistivity relation known as Archie's law
is commonly observed for rocks.
P
P0
_
-r
24
p = the resistivity of the brine-saturated rock
p
= the resistivity of the brine
n = pore porosity
r = constant (2.0 ± 0.1)
Greenberg and Brace [1969] found that networks of
resistors behaved like brine-saturated rocks.
Networks with
3 to 6 intersecting elements gave r values in the range 2 to 3.
For most granites, 2 to 4 cracks are commonly observed to
intersect.
2 to 3.
The number for Westerly granite is somewhat lower,
For granites, various upper limits have also been
placed on the number of cracks sharing an intersection (Table
3).
In Red Lake granite as many as six cracks are observed
intersecting at a common point, while for Westerly granite
the maximum number is three.
The other granites fall in
between.
Scale of Features
Many important microstructures in low porosity
crystalline rocks are a few microns or less in size.
The
extremely fine scale of the microporosity rules out detailed
crack investigations based upon ordinary thin or polished
sections.
Most of the structures revealed by the scanning
electron microscope in ion-thinned surfaces would be hidden
by polishing damage in ordinary sections.
Ordinary polishing tends to etch out grain boundaries.
Delicate bridges and slots are destroyed so that instead of
a row of slots a groove is observed.
The abrasives striate
the mineral surfaces and fill the pores with debris.
In
unthinned specimens, only the largest pores are visible.
Thin sectioning creates many transgranular cracks such as
those seen by Paulding [1965] in Westerly granite.
All of
the features we have observed in Westerly granite have lengths
less than a grain boundary.
Radioactive decoration of cracks in retrospect is
unfeasible because the grain size of the film is larger than
the features in the rocks.
Other decoration schemes would
fail to reveal the finer structures, because not enough
coloring matter can be concentrated in the microstructures.
The Concept of Cracks and Pores
Pores are approximately equidimensional cavities which
are believed to provide networks of conduction paths to very
high pressure [Brace, 1971a; Walsh, 1965a].
Cracks are
extremely narrow openings such as cleavages and grain boundary
separations which can be closed by application of a few
kilobars of pressure.
These are the definitions of cracks
and pores used in the theories explaining the various physical
properties.
In many micrographs such as Figure 2 a natural division
between cracks and pores seems to be present.
The feldspar
grain is filled with more or less equant cavities.
The quartz
grain contains a few equant cavities.
Also, the crack in
regions such as "a" intersects pores.
However, are these
really pores in the sense of the definition above?
the definition
To satisfy
the equant cavities must be connected.
Are
these equant cavities tubes of nearly circular cross-section?
Are they connected by fine cracks in another plane?
When the same field was rephotographed after about
30 ± 10pM had been removed, there was still a crack in the
same general area with approximately the same curve.
quartz, there was still a cluster of pores at "b".
pores could be tubes.
In the
Thus these
Unfortunately it was impossible to
unequivocably follow any individual pores.
The cracking observed in various stress experiments
seems to indicate that most of the pores in the feldspar
are not connected, although some of those in quartz are.
When
the same area was viewed after thermal stressing, only one new
crack appeared in the feldspar.
Presumably cracks connecting
the pores just below the surface prior to heating would have
grown when the rock was thermally stressed and would now
intersect the plane of view.
After thermal stressing, more,
but not all, of the pores in the quartz were connected.
When
rocks are mechanically stressed, the feldspars generally crack
along cleavage planes rather than preferentially through the
pores.
Further evidence arguing against the interconnectivity
of most of the pores in feldspar is that if they were connected
the porosity as measured by immersion should be higher.
As
mentioned earlier, the porosity measured 'from the micrographs
is significantly higher than that measured by immersion.
Many of the pores in the feldspar could be fluid inclusions.
What then provides the conduction paths at high pressure?
Some of the pores in quartz show evidence of being interconnected by cracks in another plane.
Figure 2 does intersect some pores.
Also, the crack in
Another consideration
is that many cracks are very irregular in cross-section and
would probably not close completely with pressure.
The concept of cracks and pores may need some
redefinition because many of the pores are not pores in the
sense of the models.
Some of the behavior attributed to
pores may come from cracks which cannot close completely.
SUGGESTIONS FOR FUTURE RESEARCH
The method used here for viewing microstructure has
enormous potential.
Other researchers have reported crack
orientations, crack lengths, crack densities, and crack
intersections.
In addition to measuring these parameters
I have recorded crack aspect ratios, crack widths, and
details of crack and pore shape for six granites, a diabase,
and a gabbro, but many rock types have yet to be examined.
28
A preliminary investigation of the effects of
mechanical stress on rocks showed that many cracks develop
parallel to the direction of maximum compression.
Determination of the onset of brittle cracking and a
thorough SEM study of the stages into which the typical
stress-strain curve has been divided [Brace, 1971b; Brace
et al.,
1966] would be very valuable.
Before cracks
preferentially grow parallel to the direction of loading,
do pre-existing cracks close, do randomly-oriented cracks
form, and does sliding occur on pre-existing surfaces as
hypothesized?
buckles.
In later stages the specimen spalls and
Steeply inclined shear fractures appear, and
finally an open fault develops.
What microscopic changes
accompany these large-scale fractures?
When Chelmsford granite was heated to 400 0 F, many new
cracks formed.
The minimum temperature change that causes
crack growth is still unknown.
Hydrostatic pressure is
widely believed to prevent thermal cracking.
How much
pressure is necessary to- balance a given temperature change?
Does hydrostatic pressure by itself cause cracking?
SEM
studies could resolve these questions.
The porosity of the feldspars as seen with the SEM is
a new puzzle.
Why does the size of the pores in the feldspars
vary from granite to granite?
Why don't the plagioclases in
Maryland diabase and San Marcos gabbro contain pore
concentrations?
Possible suggestions are that the porosity
depends on the plagioclase composition, or the water
saturation of the melt, or the pressure at which the mineral
solidified.
The use of solid cores has proved to be a very good way
to study structure, because the core can be further sectioned
or stressed.
However, the determination of the mineralogy of
a given field is difficult, because the electron microprobe
should be used to find the composition.
In many cases it
might be very advantageous to work with ion-bombarded thin
sections.
Such sections could be examined on an ordinary
microscope to determine the mineralogy, coatedand then
viewed with the SEM.
ACKNOWLEDGEMENTS
I would like to thank Professor W. F. Brace for
suggesting this project.
Frequent discussions of the
photomicrographs with him contributed to the excitement
and sense of discovery I felt in pursuing this study of
cracks.
His advice and encouragement were very valuable.
Dr. D. L. Kohlstedt's aid in arranging for my use of
the Metallurgy Department's Commonwealth Scientific ion
milling instrument was very much apureciated.
I also want
to thank him for his advice on using that machine,
Discussions of electron microscopy techniques with him
were very useful.
R. lHon was a wonderful source of petrological advice.
He also gave me the pieces of Katahdin Granite.
H. Noyes
assisted me by providing a chunk of the granodiorite of
Red Lake.
Thanks are also due to Madge Slavin who typed the
final copy.
In general I would like to say how very much
I enjoyed working with everyone on the seventh floor ofthe Green Building.
APPENDIX I
POLISHING HOLDER
Since the samples were very small, a polishing holder
was designed to grip the samples while they were ground on
the metal wheel and polished on the cloth laps.
Not only did
the holder provide an easier size object to grasp, but it
also insured that the sample surface stayed perpendicular
to its axis during the polishing and grinding.
A diagram
of the holder is shown in Fig. 9.
The rock was inserted into the center hole.
Shims made
of pieces of paper of known thickness with holes of the same
diameter as the sample were placed on top of a glass plate.
The rock and the holder were then placed on top of the shims
and the rock was tapped gently until it touched the glass
plate so that a measured amount of rock protruded from the
holder.
Then the holder's screw was tightened to grip
the rock firmly and the protruding rock was polished away.
Assuming that the abrasive produced a damaged layer directly
proportional to its size,
the process was repeated six times
with progressively finer abrasives and thinner shims.
The
polished rock surface was then ready for ion bombardment.
32
APPENDIX II
CHOICE OF COATING
Specimens viewed with the SEM must have conducting
surfaces.
In electron microscopy many substances including
gold-palladium, gold, platinum, silver, graphite, and
aluminum are vacuum evaporated onto sample surfaces to
provide a conducting layer.
I chose a 40% gold 60% palladium
alloy after experimenting with several metals.
At first gold was used, but it had two disadvantages.
The gold coating tended to crack.
Thus the approximately
.5/cm or smaller cracks might have obscured the fine scale
structures.
Also at high magnification the surface charged
and darkened so that when the magnification was reduced a
dark square was visible.
In an attempt to avoid cracking, silver was tried,
Unfortunately the silver reacted with the.air to form an
oxidized layer.
Platinum produced very good results, because it did not
crack or charge.
However, its high melting tenmperature
made vacuum evaporation more difficult.
The gold-palladium alloy did not crack and had a lower
melting point than platinum.
At high ragnification it
charge, but not nearly as fast as gold did.
did
APPENDIX III
HOLDER FOR THE ROCK
Standard specimen mounts for the Jeolco scanning
electron microscope are designed to hold .375"
diameter specimens.
or 1.00"
Fig. 10 is a diagram of the holder
for the rock, which is meant to be used in the standard
1.00" diameter mount.
Holders with different inside
diameters can be used to adapt other size cores to the
standard microscope mounts.
34
APPENDIX IV
PRECISE RELOCATION OF A PARTICULAR SEM FIELD
Several aspects of this study were based upon observing
changes in a given SEM field and thus upon the ability to
precisely relocate it.
For serial sectioning it was
necessary to return to the same field after approximately
Z3Opm of material was removed.
To study themial stress
selected fields were examined both before and after heating.
The problem was to retain the orientation of the
cylindrical sample so that the coordinate system used in the
first viewing would not be rotated.
To accomplish this.
the rock core was placed inside a holder (Apuendix III)
whose set screw fitted into the groove along one side of the
rock core.
The holder was then placed in the microscope
mount, and a groove in the side of the holder was aligned
with the screw in the mount.
be seen in Fig. 11.
The entire configuration can
The sample was then inserted in the
SEM in the same general orientation every time.
The rock's groove when viewed on end looked like a
notch, and was used as the origin of coordinates.
The
precise coordinates of the notch usually shifted slightly
from one viewing to the next.
The shift was recorded and
the new coordinates of the desired field were calculated.
The TV mode of observation at lowest magnification (lOOX)
35
was used to relocate the field using the computed coordinates
as a guide.
Once the field was found, the correct coordinates
were recorded.
The TV was turned off and the cathode ray
tube display was used for viewing because it provided better
resolution and because photographs could be taken of the
display.
APPENDIX V
MICROPROBE ANALYSIS
Although the SEM was an excellent tool with which to
study microstructure, it left many mineralogical questions
unanswered.
The SEM samples were solid cores, so the
standard petrological techniques of the optical transmission
microscope could not be used.
Instead an electron
microprobe was used to determine the composition of the
various grains and to verify the location of the grain
boundaries.
The electron microprobe had several important
limitations.
An optical reflecting microscope with a
magnification of 400X was used to select the points for
analysis.
The magnification and the shallow depth of
field of the microscope were severe handicaps.
The
microprobe beam can theoretically be reduced to lpim, but
generally a wider beam (6pkm) was used.
to a depth of 6 Lm.
This beam samples
The SEM saw only surface structures, while
the probe analysis reflected subsurface composition.
If
for example a thin piece of quartz lay on top of plagioclase,
the SEMl saw structures in quartz, where the probe gave a
high calcium count.
The microprobe study revealed that the plagioclase
grains in granite contained pore concentrations.
The
microcline in Chelmsford Granite was almost as porous as the
37
plagioclase, but the microcline in Westerly Granite had so
few pores it was frequently confused with quartz, which
typically has only a few scattered pores.
Traverses across several grains (Fig. 12) were made in
an attempt to verify whether or not we were correctly
identifying grain boundaries with the SEiM.
The raw data
in Fig. 12 implies that the grain was thinner towards its
edges.
The microprobe data concurred with the SEM
photomicrographs.
IOOX
to
10,OOOX
Unfortunately features observcd at
with the SEM were invisible to the
microprobe microscope (400X),
so it
was impossible to
verify whether or not they lay precisely at the grain
boundary.
38
BIBLIOGRAPHY
Adams, L.H., and E.D. Williamson, The compressibility of
minerals and rocks at high pressure, J. Franklin. Inst.,
195, 475-529, 1923.
Baldridge, S., and G. Simmons, Progress in microcrack
decoration, EOS, 52, 342,
1971.
Birch, F., The velocity of compressional waves in rocks to
10 kb, I, J. Geophys. Res., 65, 1083, 1960.
Brace, W.F., Dependence of fracture strength of rocks on grain
size, Penn. State Univ. Mineral Industries Expt. Sta. Bull.
No. 76, 99-103, 1961.
Brace, W.F., Brittle fracture of rocks in State of Stress in
the Earth's Crust, ed. by W.R. Judd, pp. 110-178, Am.
Elsevier Pub. Co., Inc., 1964.
Brace, W.F., Some new measurements of linear compressibility of
rocks, J. Geophys. Res., 70, 391-398, 1965.
Brace, W.F., B.W. Paulding, and C. Scholz, Dilatancy in the
fracture of crystalline rocks, J. Geophys. Res.,
71, 3939-
3953, 1966.
Brace, W.F., and A.S. Orange, Further studies of the effects of
pressure on electrical resistivity of rocks, J. Geophys.
Res., 73, 5407-5420, 1968.
Brace, W.F., J.B. Walsh, and W.T. Frangos, Permeability of
granite under high pressure, J. Geophys. Res.,
73,
2225-2236, 1968.
Brace, W.F., Resistivity of saturated crustal rocks to 40 km
based on laboratory measurements, Geophysical Monograph
Series, Vol. 14,
The Structure and Physical Properties
of the Earth's Crust, ed. by John G. Heacock, pp. 243-255,
Amer. Geophys.Un., Washington, D.C.,
1971a.
Brace, W.F., Micromechanics in rocks systems, Proc., Int. Conf.
Structure, Solid Mechanics and Engineering Design, Univ.
Southampton, England
(1969),
John Wiley & Sons, N.Y.,
Brace, W.F.,
ed. M. Te'eni, pp. 187-204,
1971b.
Pore pressure in geophysics, Geophysical
Monograph No. 16, Flow and Fracture of Rocks, pp. 265-274,
Amer. Geophys. Un., Washington, D.C.,
Brace, W.F.,
1972.
E. Silver, K. Hadley, and C. Goetze, Cracks and
pores: A closer look, Science, 178, 162-164, 1972.
Greenberg, R.J.,
and W.F. Brace, Archie's Law for rocks
modeled by simple networks, J. Geophys. Res.,
2102,
Kimoto, S.,
74,
1969.
and J.C. Russ, The characteristics and applications
of the scanning electron microscope, Amer. Sci.,
1969.
2099-
57,
112,
Koide, H.,. and K. Hoshino, Development of microfractures in
experimentally deformed rocks (Preliminary Report),
Jishin, 20, 85-97, 1967.
Majumdar, A.J., The application of scanning electron microscopy
to textural studies, Proc. British Ceramic Soc., No. 20,
pp. 43-69, 1970.
Nur, A.,
and G. Simmons, The effect of saturation on velocity
in low porosity rocks, Earth & Planet. Sci. Letters, 7,
183-193, 1969.
Oatley, C.W., W.C. Nixon, and R.F.W. Pease, Scanning electron
microscopy in Advances in Electronics and Electron Physics,
Vol. 21, pp. 181-247, Academic, N.Y., 1965.
Orange, A.S., Granitic rock: Properties in situ, Science, 1965,
202-203, 1969.
Paulding, B.W., Jr., Crack growth during brittle fracture in
compression, PhD Thesis, Mass. Institute of Technology, 1965.
Simmons, G., Velocity of compressional waves in various minerals
at pressures to 10 kilobars, J. Geophys. Res., 69, 1117-1121,
1964.
Timur, A., W.B. Hempkins, and R.M. Weinbrandt, Scanning electron
microscope study of pore systems in rocks, J. Geophys. Res.,
76, 4932-4948, 1971.
41
Todd, T.,
and G. Simmons, Effect of pore pressure on the
velocity of compressional waves in low porosity rocks,
J. Geophys. Res.,
77, 3731-3743, 1972.
Walsh, J.B., The effect of cracks in rocks on Poisson's ratio,
J. Geophys. Res., 70, 5249-5257, 1965a.
Walsh, J.B., The effect of cracks on the compressibility of
rock, J. Geophys. Res., 70, 381-389, 1965b.
Walsh, J.B., The effect of cracks on the uniaxial elastic
compression of rocks, J. Geophys. Res., 70, 399-411, 1965c.
Walsh, J.B., and W.F. Brace, Elasticity of rock: A review of
some recent theoretical studies, Rock Mech. & Engrg.
Geol., 4, 284-297, 1966.
Walsh, J.B., and E.R. Decker, Effect of pressure and saturating
fluid on the thermal conductivity of compact rock, J. Geophys.
Res., 71, 3053-3061, 1966.
Wawersik, W.R., and W.F. Brace, Post-failure behavior of granite
and diabase, Rock Mechanics, 3, 61-85, 1971.
White, E.W., and R. Roy, Applications of the scanning electron
microscope to ceramic studies, Proc. of Sympos. The Scanning
Electron Microscope: The Instrument and Its Applications,
ITT Res. Inst., Chicago, pp. 91-94, 1968.
42
Willard, R.J., and J.R. McWilliams, Microstructural techniques
in the study of physical properties of rocks, Int. J. Rock
Mech. Min. Sci., 6, 1-12, 1969.
43
LIST OF ILLUSTRATIONS
Figure 1.
Photomicrograph of intersecting cracks within a
quartz grain in Westerly Granite
2.
Photomicrograph of Chelmsford Granite
3.
Photomicrograph of a line of pores in a plagioclase
grain in San Marcos Gabbro.
4.
Photomicrograph of a multigrain field in
Westerly Granite.
5.
Photomicrograph of Westerly Granite (#18) which
was stressed in triaxial compression to over
95% of its fracture strength.
6.
7A.
Photomicrograph of Westerly Granite (#18).
Composite of three photomicrographs of stressed
Westerly Granite (#18).
7B.
Map of the cracks in the field of 7A.
70.
Histogram of the crack orientations with respect
to the direction of maximum compression.
8.
Photomicrograph of Maryland Diabase stressed in
triaxial compression to over 95% of its fracture
strength.
9.
Polishing Holder.
10.
Holder for Rock.
11.
Sample ready for insertion into the SEM.
1 .
Electon microprobe study of quartz-plagioclase
grain boundaries in Chelmsford Granite.
LIST OF TABIES
Table I.
II.
III.
Rocks Studied.
Stressed Rocks Studied.
Comparison of Observations to Predictions.
Fig. 1:
Photomicrograph of intersecting cracks within a
quartz grain in Westerly Granite.
interiors and white edges.
irregular gray areas.
The cracks have black
The thinning mounds are the
*A,4'-
w
A
qV
Figure 1
Fig. 2:
Photomicrograph of Chelmsford Granite.
The highly
porous grain to the left is plagioclase (pg), while the
grain to the right is quartz (qtz).
The bright band running
down the middle is the topographic step at the grain
boundary, caused by the difference in thinning rates
between the grains.
A crack runs along the grain boundary
until halfway up the picture, it turns and crosses the
quartz grain above "b".
A3
On
-4
r4
rz
47
Fig. 3:
Photomicrograph of a line of pores in a plagioclase
(pg) grain in San Marcos Gabbro.
left corner is an amphibole.
crack free.
The grain in the bottom
The grain boundary is
oil
Figure 3
48
Fig. 4:
Granite.
Photomicrograph of a multigrain field in Westerly
The center quartz (qtz) grain is surrounded by
microcline (mi), biotite (bi) and plagioclase (pg) grains.
The bright areas at "a",
"b",
and "c" are cracks which
occupy about 20% of the quartz grain boundary.
irregular mounds are from the ion bombardment.
The
*
*
1'~
*
~
I
~~Jk
Fig. 5:
Photomicrograph of Westerly Granite (#18) which
was stressed in triaxial compression to over 95% of its
fracture strength.
maximum compression.
The white arrows show the direction of
1,
0T
-I
L()
a)
-el
E-T-1
50
Fig. 6:
Photomicrograph of Westerly Granite (#18).
The
white arrows show the direction of maximum compression.
The quartz (qtz)-plagioclase (pg) boundary runs from
top to bottom to the right of "c".
eTFO
Y
/
ti
3R-1
Fig. 7A:
Composite of three photomicrographs of stressed
Westerly Granite (#18).
The white arrows show the
direction of maximum compression.
Fig. 7B:
Map of the cracks in the field of 7A.
Fig. 7C:
Histogram of the crack orientations with respect
to the direction of' maximum compression.
The angular
distribution of the cracks in 7A is shown by the solid
line while the dashed line shows the distribution for
aracks in a similar field of the same size in unstressed
Westerly Granite.
Fiqure 7A
Figure 7B
Crack Orientation with respect
to the direction of Ma>imum
Compression
50stressed
unstressed
40
0 Ln
32130
0(Om
4---
O 210.,
1~-*'
.1-----
I
---.
~
*
.L
4.-
-J.
0
20
Angle, degrees
Figure 7C
8O
52
Fig. 8:
Photomicrograph of Maryland Diabase stressed in
triaxial compression to over 95% of its fracture strength.
The pyroxene (pyr)-plagioclase (pg) grain boundary is the
site of a large topographic step because of the difference
in the minerals' thinning rates, but the grain boundary
does not contain cracks.
In the center of the micrograph
a crack crosses the grain boundary.
Ql
Ll
41V
.4~~
~~.-1
Fiqure 8
Fig. 9:
Polishing Holder.
please see Appendix I.
For details on its use
POLISHING
HOLDER
42" D01"
It.
-
1.5"t. 2
Figure
plIxigIlass
9
Fig. 10:
Holder for Rock.
please see Appendix III.
For details on its use
FOR ROCK
HOLDER
.g8t .01"
42't1'l
-A
(p
I4-~
9
brass
Figure 10
Fig. 11:
Sample ready for insertion into the SFR.
rock
Figure 11
56
Fig.. 12:- Electron microprobe study of quartz-plagioclase
grain boundaries in Chelmsford Granite.
CO crack
P:
QUARTZ
path
1
-
-QUARTZ
.2
RAW MICROPROBE DATA:
Note:
Point
Potassium
Calcium
1
.00
.02
2
.00
.00
3
.01
.00
.02
.29
.05
1.13
.05
.85
.00
.07
.00
.00
The points were roughly 18/4m apart.
TABLE I
Rocks Studied
Crack
Rock
Porosity (%)
Pore
Total
Grain
Size
3
(10'-mm)
Diabase I
Frederick, Md.
0.1
0.1
180
Gabbro
San Marcos, Cal.
Raymond granite
Madera, Cal.
0.2
0.2
.22
0.4
Katahdin granite I
Baxter State Park,
Maine
.20
Chelmsford granite
Massachusetts
Westerly granite
Rhode Island
Modal
Analysis
Reference
48 an6 7 , 49 pyr,
1 mica, 1 ol
Brace & Orange, 1968
Birch, 1960
2000
70 an4 2 , 12 mica,
8 pyr, 7 am, 3 ox
Brace & Orange, 1968
0.6
2000
47 ani 6 - 20 , 35 qu,
10 mi, 7 mica
Sprunt (present study)
0.7
0.9
2000
28 ani 8 - 2 4 , 35 qu,
27 mi, 9 mica
Sprunt (present study)
.36
0.7
1.0
1500
Birch, 1960
Todd & Simmons, 1972
.2
1.2
1.1
31 qu, 31 mi,
31 an 2 0 - 2 4 , 7 mica
27.5 qu, 35.4 mi,
31.4 an1 7 , 4.9 mica
750
Brace, 1965, 1964a
Birch,
1960
Simmons, 1964
Granodiorite of
Red Lake
Shaver Lake, Cal.
.29
0.8
1.1
2000
49 an 3 0 , 27 mi,
17 qu, 5 mica,
2 misc
Sprunt (present study)
Katahdin granite II
Baxter State Park,
Maine
.46
0.7
1.2
3000
29 ani 8s-2 4, 29 qu,
33 mi, 9 bi
Sprunt (present study)
an = plagioclase
ox = oxide
pyr = pyroxene
qu = quartz
ol = olivine
am = amphibole
bi = biotite
mi = microcline
TABLE
II
Stressed Rocks Studied
Rock
Confining
Pressure
Strength
Reference
Westerly granite #18
1.5 kb
10.52 kb
Paulding, 1965
Westerly granite #28
0.5 kb
6.30 kb
Paulding, 1965
Maryland diabase
1.6 kb
13.21 kb
Wawersik & Brace,
1971
TABLE III
Comparison of Observations to Predictions
Property
Rock
Observed (SEM)
Predicted
Order of the
grain size
Reference
Brace & Walsh, 1966
Brace, 1961
Brace, 1964a
Crack length
Granites
Less than a grain
diameter, usually
200-400 ym
Aspect ratio
Granites
-10
-103
Walsh, 1965a
Brace et al., 1968
Crack porosity
Chelmsford granite
.26%
0.29%
Todd, personal
communication
Maryland diabase
.05%
0
Brace & Orange, 1968
Brace, 1965
1.2%
0.7%
Todd, personal
communication
2 to 4
2 to 6
Pore porosity
Chelmsford granite
Number of cracks Granites
intersecting at
a point
Greenberg & Brace, 1969